The use of discriminant analysis for the interpretation of the reflectancespectra of irradiated porcine M-longissimus dorsi

Slices of porcine M. longissimus dorsi were packed in overwrap packs and su bjected to irradiation (0 and 5 kGy) and then stored for 7 days at 4 degree s C. Reflectance spectra were measured on the outside surface and a freshly cut surface at 7 days post irradiation. The reflectance spectra were trans formed to reflex attenuance, k/s and first and second difference spectra an d subjected to discriminant analysis. Using discriminant analysis it was po ssible to establish a calibration equation to discriminate between the spec tra of irradiated and unirradiated pork for both the outside and the inside surface. When the calibration model was used to predict the classification of new samples a 100% correct grouping was obtained for the freshly cut su rface, however, for the outside surface the classification ranged from 87 t o 100% correct depending on the mathematical transformation of the reflecta nce spectra. This shows the potential of colour measurements as a possible rapid initial screening test for the identification of irradiated pork. Eva luation of the first difference spectra to identify peak positions showed s ignificant differences in peak positions between irradiated and unirradiate d pork. The position of the peaks in the irradiated sample is discussed as lending support to the hypothesis of the carboxyhaem form as the irradiated pigment. (C) 2000 Elsevier Science Ltd. All rights reserved.